Five mammalian isoforms of the G protein β subunit (37 kDa) and twelve isoforms of G protein Y (7.8 kDa) have been identified (Offermanns (2003) Prog. Biophys. MoI. Biol 83:101-30). Obligate heterodimers composed of G protein β and Y subunits (Gβγ) function as regulatory molecules in various pathways in eukaryotic cells (Neves, et al. (2002) Science 296:1636-9; Clapham and Neer (1997) Annu. Rev. Pharmacol. Toxicol. 37:167-203). First characterized as a guanine nucleotide dissociation inhibitor (GDI) , Gβγ associates tightly with GDP-bound G protein α subunits (Ga) and thereby constitutes the basal form of the G protein heterotrimer in which neither Ga nor GβY are active in signaling. Agonist-stimulated G protein coupled receptors (GPCRs) catalyze the exchange of GDP for GTP upon Ga and release of Gβγ from the heterotrimer complex, liberating two active signaling species: Gaα>>GTP and Gβγ. Targets of Gβγ signaling include the G protein-regulated inward-rectifying potassium channel (GIRK) (Krapivinsky, et al. (1993) J. Biol. Chem. 273:16946-52); type I, type II, and type IV isoforms of adenylyl cyclase (Tang and Gilman(1991) Science 254:1500-3; Sunahara, et al. (1996) Annu. Rev. Pharmacol. Toxicol. 36:461-80); mitogen-activated protein kinase (MAPK) (Schwindinger and Robishaw (2001) Oncogene 20:1653-60); phosphotidylinositol-3-kinase (PI3K) (Schwindinger and Robishaw (2001) supra); phosducin (Schulz (2001) Pharmacol Res 43:1-10); at least two members of the G protein receptor kinase (GRK) family (Koch, et al. (1993) J. Biol. Chem. 268:8256-60; Inglese, et al. (1994) Proc. Natl. Acad. Sci. USA 91:3637-41); and other plextrinhomology (PH) domain-containing proteins including the dynamins (Lin, et al. (1998) Proc. Natl. Acad. Sci. USA 95:5057-60; Scaife and Margolis (1997) Cell Signal 9:395-401) and the β1, β2, and β3 isoforms of phospholipase C β (PLC β) (Sternweis and Smrcka (1992) Trends Biochem. Sci. 17:502-6; Li, et al. (1998) J. Biol. Chem. 273:16265-72) and many others.
Gβ is a cone-shaped toroidal structure composed of seven four-stranded β-sheets arranged radially about a central axis (Wall, et al. (1995) Cell 83:1047-58; Lambright, et al. (1996) Nature 379:311-9). Each β-sheet is formed from elements of two consecutive WD-40 repeats, named for a conserved C-terminal Trp-Asp sequence in each repeat (Gettemans, et al. (2003) Sci STKE 2003:PE27). The GY subunit, an extended helical molecule, is nested in a hydrophobic channel that runs across the base of the cone. The slightly narrower, “top” surface of the Gβ cone is the main binding site of Ga (through its switch II region) (Wall, et al. (1995) supra; Lambright, et al. (1996) supra), phosducin (Loew, et al. (1998) Structure 6:1007-19; Gaudet, et al. (1996) Cell 87:577-88), and GRK2 (Lodowski, et al. (2003) Science 300:1256-62), as shown by the crystal structures of these complexes. Mutational analysis indicates that many interaction partners of Gβy, including PLC β2 and adenylyl cyclase, bind to the same surface (Li, et al. (1998) supra; Ford, et al. (1998) Science 280:1271-4). Sites located along the sides of the Gβ torus serve as auxiliary binding surfaces that are specifically recognized by certain Gβy targets, exemplified in the crystal structures of Ga and phosducin bound to Gβy (Wall, et al. (1995) supra; Loew, et al. (1998) supra; Gaudet, et al. (1996) supra; Wall, et al. (1998) Structure 6:1169-83).
Phage display of randomized peptide libraries has been used to identify sequence requirements for binding and screen for peptide that bind to Gβ1γ2 dimers (Scott, et al.(2001) EMBO J. 20:767-76). Although billions of individual clones were screened, most of the peptides that bound Gβχγ2 could be classified into four, unrelated groups based on amino acid sequence. One of these groups included a linear peptide (the ΛΛSIRK″ peptide) with the sequence Ser-Ile-Arg-Lys-Ala-Leu-Asn-Ile-Leu-Gly-Tyr-Pro-Asp-Tyr-Asp (SEQ ID N0:1). The SIRK peptide inhibited PLC β2 activation by Gβ1γ2 subunits with an IC50 of 5 μM and blocked activation of PI3K. In contrast, the SIRK peptide had little or no effect on Gβ1γ2 regulation of type I adenylyl cyclase or voltage-gated N-type Ca++ channel activity (Scott, et al. (2001) supra). This demonstrated that selective inhibition of Gβγ binding partners could be achieved. Peptides belonging to all four groups competed with each other with a range of affinities for binding to Gβ1γ2, suggesting that all of the clones isolated from the phage display screen shared a common binding site on Gβ1γ2 (Scott, et al. (2001) supra).
Subsequent experiments have shown that not only does the SIRK peptide block heterotrimer formation, but it also displaces Gαn from a Gβ1γ2>>Gαii complex in the absence of Gαii activation and activates G protein-dependent ERK1 and ERK2 pathways in intact cells (Ghosh, et al. (2003) J. Biol. Chem. 278:34747-50; Goubaeva, et al. (2003) J. Biol. Chem. 278:19634-41). In vitro experiments revealed that SIRK facilitated nucleotide exchange-independent-A-heterotrimer dissociation (Goubaeva, et al. (2003) supra; Ghosh, et al. (2003) supra) potentially explaining the activation of ERK in intact cells. Other Gβγ binding peptides such as QEHA, derived from adenylyl cyclase II (Weng, et al. (1996) J. Biol. Chem. 271:26445-26448; Chen, et al. (1997) Proc. Natl. Acad. Sci. USA 94:2711-2714) and amino acids 643-670 from the C-terminal region of βARK(GRK2) (Koch, et al. (1993) supra) could not promote dissociation of the heterotrimer, despite competing for Ga subunit binding (Ghosh, et al. (2003) supra). This indicates that competition for Gα-Gβγ subunit binding is not sufficient for these peptides to accelerate subunit dissociation.
Using a doping mutagenesis and rescreening strategy, a peptide similar to the SIRK peptide was derived that had higher affinity for Gβ1γ2. The sequence of this peptide is Ser-Ile-Gly-Lys-Ala-Phe-Lys-Ile-Leu-Gly-Tyr-Pro-Asp-Tyr-Asp (SEQ ID N0:2) (SIGK). In vitro studies with the SIGK peptide indicate that it too can displace Gαn from a heterotrimeric complex and also effectively prevents heterotrimer formation (Ghosh, et al. (2003) supra). The mechanism by which SIRK/SIGK mediates the dissociation of Gαii* GDP from Gβ1γ2 is not fully understood (Ghosh, et al. (2003) supra).